Pacman

Historical bug tracker for the Pacman package manager.

The pacman bug tracker has moved to gitlab:
https://gitlab.archlinux.org/pacman/pacman/-/issues

This tracker remains open for interaction with historical bugs during the transition period. Any new bugs reports will be closed without further action.
Tasklist

FS#6636 - pacman -Sd should be permanent

Attached to Project: Pacman
Opened by Scott H (stonecrest) - Monday, 19 March 2007, 05:39 GMT
Last edited by Dan McGee (toofishes) - Sunday, 17 February 2008, 19:53 GMT
Task Type Feature Request
Category
Status Closed
Assigned To Dan McGee (toofishes)
Architecture not specified
Severity Medium
Priority Normal
Reported Version 0.7.2 Gimmick
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

When a user installs a package with the -d flag, it seems that this should be a permanent action. Meaning that when the package updates in the future, the -d flag should continue to hold rather than having all its deps pulled in. Presumably if you didn't use dep-checking once, you wouldn't want it again.
This task depends upon

Closed by  Dan McGee (toofishes)
Sunday, 17 February 2008, 19:53 GMT
Reason for closing:  Won't implement
Additional comments about closing:  Too complex for pacman and could lead to problems for users.
Comment by Dan McGee (toofishes) - Tuesday, 20 March 2007, 17:00 GMT
This fits in with your concerns about --noscriptlet as well- if one installs a package without using the scriptlet, it should probably be removed and upgraded without it.
Comment by Scott H (stonecrest) - Tuesday, 20 March 2007, 17:20 GMT
I have to admit that I am having some uneasiness about this now. But I think it's due to the fact that the current pacman 2.9.8 sometimes forces users to use -Rd (e.g. http://www.archlinux.org/news/201/) in order to get around pacman bugs. Hopefully with pacman 3, users will only use -Rd for their own reasons (1), not because they need to work around pacman bugs. But I guess, bottom line, is that folks should only be using -Rd if they know what they're doing.


1) I use pacman -Rd with gnome-python-extras because the systemtray support works perfectly without having to install a huge number of gnome deps.
Comment by Hussam Al-Tayeb (hussam) - Monday, 26 March 2007, 23:27 GMT
Sorry for the bug spam but making pacman -Sd permanent is a bad idea in my opinion. It could eventually cause breakage on a user's system.
Comment by Scott H (stonecrest) - Tuesday, 27 March 2007, 16:40 GMT
(Oops, replace all -Rd with -Sd in my previous message :P)

Sure it could eventually cause breakage, but using pacman -Sd can cause breakage the very first time you use it too. If someone uses -Sd, they should know what they are doing (assuming, again, that they aren't doing pacman workarounds). And if you're not working around a pacman bug, what's the point of using -Sd when, on your next -Syu, it's going to retrieve all of a package's dependencies?
Comment by Roman Kyrylych (Romashka) - Tuesday, 27 March 2007, 17:30 GMT
1) I think -Sd will become less usefull when packages won't depend on packages that are optional dependencies, this will require package maintainers to sanitize depends field in their packages and use optdepends (or whatever it will be called) feature (http://www.archlinux.org/pipermail/pacman-dev/2006-October/000577.html and http://bugs.archlinux.org/task/4845#comment10357).

2) Also same result as permanent -Sd could be reached with improved IgnorePkg (http://bugs.archlinux.org/task/6430).

3)Nevertheless _automatic_ permanent -Sd can be implemented by putting '!' before dependency in /var/lib/pacman/local/<pkgname>/depends, i.e.:
%DEPENDS%
foo
!bar
In this way pacman won't lose track of dependencies while not installing them again and again on -Syu.
But will it be worthy to implement this for few very specific uses when there will be 1st & 2nd (which are useful on their own)?
Comment by Scott H (stonecrest) - Tuesday, 27 March 2007, 19:52 GMT
In my specific case, 1 or 2 would not be very helpful. For example, I choose to pacman -Sd gnome-python-extras because, for my very specialized use case of only wanting egg systemtray support, all of the other gnome deps are not needed. It wouldn't make sense for the gnome deps to become optional, and there would be a lot of packages i would need to add to IgnorePkg to have this work correctly.

Frankly, the easiest thing right now is to pacman -Sd gnome-python-extras each time that I see it updated.
Comment by Xavier (shining) - Sunday, 29 July 2007, 18:26 GMT
I don't really like the idea of a permanent -Sd either. Btw, how would you override it, if you want dependencies checking back for a package previously installed with -Sd ?
Anyway, this looks like a wrong solution.
The idea of optional dependencies seems to be perfectly adapted to this use case.
But if in this specific case, the optional dependencies are not really optional, then the sanest thing to do here is making a second package imo.

Or maybe a smarter optional deps system could handle this also. For example, some optional deps could be installed by default at the first installation
of the package. And there could also be a way to specify exactly which optional deps to install.
Comment by Dan McGee (toofishes) - Sunday, 17 February 2008, 01:29 GMT
Can we close this as a "Won't implement"? Most of the feedback associated with it has said this is probably not necessary, and a power-user could figure out that they either need to (1) use -Sd to upgrade the package (2) build their own version of the pkg without the 'offending' deps.
Comment by Xavier (shining) - Sunday, 17 February 2008, 08:31 GMT
As my comment suggested, I'm for the 'won't implement'.

Loading...